prestashop.com — AI Site Grade

Analysis

I have enough data to write a thorough audit. Let me compile the findings.

PrestaShop.com: AI crawlers get full access but find a schema desert and a missing llms.txt

Every major AI crawler — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended — receives a 200 with the full 302KB HTML payload from Cloudflare, identical to a browser. No UA-based blocking, no JS shell, no 403. Yet the site's structured data is the thinnest possible: every page carries only Organization and WebSite schema with name and URL — no Product, SoftwareApplication, FAQPage, BreadcrumbList, Review, or HowTo markup anywhere on the homepage, offers page, comparison pages, or blog.

Crawler Access

The robots.txt uses a single User-agent: * catch-all with no AI-specific directives. GPTBot, ClaudeBot, Google-Extended, and others are not mentioned. The disallow list blocks WordPress admin paths, forums faceted URLs, and profile pages — standard hygiene. The llms.txt endpoint returns a 404 (serving the full 120KB Vue.js SPA shell as a 404 page). No AI-friendly content map exists. The site runs on Cloudflare behind AWS CloudFront, with no X-Robots-Tag or noindex meta on any page checked. The sitemap index is well-formed with two sub-sitemaps (pages + blog, ~550 URLs total).

Cold-Knowledge Gap

The LLM cold-knowledge snapshot describes PrestaShop as an open-source e-commerce platform founded in 2007, powering 300,000+ stores, with a strong European SMB base, mixed reviews on performance and security, and a 2023 shift to PrestaShop 8. The site itself claims 250,000 merchants (not 300,000), promotes PrestaShop 9.1 as the current version (not 8.x), and heavily pushes three commercial tiers (Classic free, Hosted, Enterprise) alongside the open-source project. The cold model knows nothing about the Hosted or Enterprise SaaS offerings, the PrestaShop Essentials suite (Checkout, Marketing, Shipping, Social, Automation), or the payment-solution partnerships (Stripe, Klarna, Alma, Airwallex). The model's "PrestaShop Cloud/Ready" reference is outdated — the site now calls this "Hosted."

Schema Posture

Across all pages examined (homepage, offers, discover, about, blog, 4 comparison pages, PrestaShop 9 page), the JSON-LD is identical: Organization + WebSite with only name and url properties. No SoftwareApplication schema for the downloadable open-source platform. No Product schema for the Hosted or Enterprise plans. No FAQPage schema on pages with FAQ sections (the discover page has 3 FAQ items, the vs-Shopify page has 5, the vs-Magento page has FAQ content). No BreadcrumbList despite deep navigation. No Review or AggregateRating schema despite the site displaying customer testimonials and "Find out what our customers are saying" sections. No ComparisonTable or ItemList on the comparison pages. This is the single largest AI-visibility gap: an AI model crawling the site cannot extract pricing tiers, feature comparisons, software version info, or FAQ answers as structured data.

Content Signals

The homepage has 311 words of visible text — thin for a category leader. The offers page and comparison pages are richer (1,500-2,000 words) but suffer from extreme navigation repetition: the full mega-menu HTML is duplicated before every page's main content, meaning an AI crawler must wade through hundreds of repeated navigation links before reaching unique body text. The blog is active and well-categorized (News, Success Stories, Tech, Marketing, Conversion, Legal, Logistics) with posts dated as recently as April 2026. The comparison pages (vs Shopify, WooCommerce, Magento, OpenCart) use FAQ sections and comparison language but lack structured comparison tables in the HTML — the comparisons appear to be rendered visually but not as semantic <table> or ItemList elements.

External Signals

DNS records reveal integrations with HubSpot, Segment, Firebase, Zendesk, Stripe, and an anthropic-domain-verification TXT record — confirming PrestaShop has explicitly verified its domain with Anthropic for Claude crawler access. The site links to an active community Slack, forums, events calendar, and a prestashop-project.org subdomain for the open-source project. The blog references a "PrestaShop Monthly" newsletter and an "important message from PrestaShop's leadership" — suggesting ongoing corporate communication that AI models would not surface without live retrieval.

Findings

1. No SoftwareApplication schema for the open-source platform High

   The site's JSON-LD only includes Organization and WebSite schema. The downloadable open-source e-commerce platform is not marked up with SoftwareApplication schema, preventing AI crawlers from extracting version, feature, or download information.

   What to change: Add SoftwareApplication schema to the homepage, open-source page, and PrestaShop 9 page with properties for name, version, operatingSystem, applicationCategory, and offers.

2. No Product schema for commercial tiers (Hosted, Enterprise) High

   The offers page lists three pricing tiers (Classic, Hosted, Enterprise) but none are marked up with Product schema. AI crawlers cannot extract pricing, features, or plan comparisons as structured data.

   What to change: Add Product schema with offers, price, and description for each tier on the offers page.

3. FAQ sections lack FAQPage schema High

   Multiple pages contain FAQ-style Q&A content (discover page, comparison pages) but none use FAQPage schema. This content is invisible to AI crawlers as structured data.

   What to change: Add FAQPage schema with mainEntity Question/Answer pairs to all pages containing FAQ sections.

4. llms.txt endpoint returns 404 High

   The standard AI-friendly content map at /llms.txt returns a 404 error (serving the Vue.js SPA shell). AI crawlers cannot discover a curated content index.

   What to change: Create an llms.txt file listing key pages (homepage, offers, open-source, blog, comparison pages) with brief descriptions.

5. No BreadcrumbList schema on any page Medium

   Despite deep navigation paths, no page includes BreadcrumbList schema. AI crawlers cannot understand site hierarchy from structured data.

   What to change: Add BreadcrumbList schema to all pages with breadcrumb navigation.

6. Customer testimonials lack Review schema Medium

   The site displays customer testimonials and 'Find out what our customers are saying' sections but no Review or AggregateRating schema is present. AI crawlers cannot extract social proof as structured data.

   What to change: Add Review schema to testimonials and AggregateRating schema where overall ratings are displayed.

7. Comparison pages lack structured comparison markup Medium

   The vs-Shopify, vs-WooCommerce, vs-Magento, and vs-OpenCart pages use visual comparison elements but no semantic HTML tables or ItemList schema. AI crawlers cannot extract feature-by-feature comparisons.

   What to change: Add ItemList or Table schema to comparison pages, or use semantic HTML tables with appropriate schema.

8. Homepage has only 311 words of visible text Medium

   The homepage contains only 311 words of unique body text, which is thin for a category-leading brand. AI crawlers have limited content to index from the primary landing page.

   What to change: Expand homepage body text to at least 800 words covering key value propositions, features, and calls to action.

9. Repetitive mega-menu HTML dilutes content-to-code ratio Low

   Every page includes the full mega-menu HTML before unique body content, forcing AI crawlers to parse hundreds of repeated navigation links. This reduces the effective content density for AI extraction.

   What to change: Consider lazy-loading the mega-menu or using a more compact navigation structure to improve content-to-code ratio.

10. LLM cold knowledge is outdated on version and offerings Medium

    The LLM cold-knowledge snapshot describes PrestaShop 8.x as current, but the site promotes PrestaShop 9.1. The cold model also lacks awareness of Hosted and Enterprise SaaS tiers, the Essentials suite, and payment partnerships.

    What to change: Ensure the site's structured data and llms.txt provide current version and offering information to correct outdated AI knowledge.

11. Robots.txt does not explicitly allow AI crawlers Low

    The robots.txt uses a catch-all User-agent: * with no specific directives for GPTBot, ClaudeBot, or other AI crawlers. While they are not blocked, the lack of explicit allowance may cause cautious crawlers to defer.

    What to change: Add explicit Allow directives for GPTBot, ClaudeBot, Google-Extended, and other AI crawlers to signal welcome access.

What's working

• All major AI crawlers receive full HTML content — Every tested AI crawler (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, Bytespider, Applebot-Extended) receives a 200 response with the full 302KB HTML payload, identical to a browser. No UA-based blocking or JS shell is present.
• Domain verified with Anthropic for Claude crawler — DNS records include an anthropic-domain-verification TXT record, confirming PrestaShop has explicitly verified its domain with Anthropic for enhanced Claude crawler access.
• Sitemap index is well-formed with two sub-sitemaps — The sitemap index at /sitemap.xml is valid and contains two sub-sitemaps (pages and blog) covering approximately 550 URLs, ensuring AI crawlers can discover all indexed content.
• Active blog with recent posts and clear categorization — The blog is regularly updated (posts as recent as April 2026) and organized into categories (News, Success Stories, Tech, Marketing, etc.), providing fresh content for AI crawlers.
• Comparison pages provide detailed feature comparisons — The vs-Shopify, vs-WooCommerce, vs-Magento, and vs-OpenCart pages each contain 1,700-2,000 words of comparison content, including FAQ sections, which AI crawlers can index as plain text.
• No noindex or X-Robots-Tag blocking on key pages — All checked pages return 200 with no noindex meta tag or X-Robots-Tag header, ensuring AI crawlers can index the content without restrictions.
• Robots.txt blocks only admin and faceted URLs — The robots.txt disallows only WordPress admin paths, forums faceted URLs, and profile pages — standard hygiene that does not interfere with AI crawler access to meaningful content.
• Site served via Cloudflare CDN for fast global delivery — The site uses Cloudflare (behind AWS CloudFront) for content delivery, ensuring low-latency responses for AI crawlers worldwide.